Skip to main content

A new continuous hybrid block method with one optimal intrastep point through interpolation and collocation


Implicit block approaches are used by a number of numerical analyzers to model mild, medium, and hard differential systems. Their excellent stability characteristics, self-starting nature, quick convergence, and large decrease in computing cost all contribute to their widespread application. With these numerical benefits in mind, a new one-step implicit block method with three intrastep grid points has been created. The major term of the local truncation error is minimized to determine which of these points is optimal. The reformulation of the suggested technique leads to a significant decrease in computing cost while maintaining the same consistency, zero-stability, \(\mathcal{A}\)-stability, and convergence. Several sorts of error are calculated, together with CPU time and efficiency plot, to determine which is superior. Differential models from the fields of heat transfer, population dynamics, and chemical engineering show that the suggested method does a better job than some of the current hybrid block and implicit Radau methods with similar properties.

1 Introduction

Ordinary differential equations are widely used in the real world, so it is crucial to develop efficient and accurate numerical methods for solving them. Several numerical techniques are proposed in this regard in the existing literature. Many domains, such as chemistry, flame propagation, computational fluid dynamics, population dynamics, engineering, and mathematical biology, require approximate solutions to tough problems, but most of these methods fall short. For models to be stiff enough, they need numerical approaches that cost a lot to compute and have unbounded stability areas. Most of the traditional methods such as explicit Runge–Kutta [1], Lobatto family [2], multi-step Adams family [35], and higher-order multiderivative types [6, 7] are not used due to either a large amount of computational effort required (very small step-size \(\Delta x \approx 0\)) or finite stability region (conditional stability). The implicit block methods, on the other hand, are preferred since they are self-starting, computationally robust (cost-effective), highly accurate, fast convergent, and \(\mathcal{A}\)-stable (a favorable property for stiff models). Some of the major advantages of the block methods include their self-starting nature and ability to overcome the overlapping of solution pieces.

We consider an initial value problem

$$\begin{aligned} \frac {dy(x)}{dx} = f\bigl(x,y(x)\bigr),\quad y(x_{0}) = y_{0}, x \in [x_{0}, T], y(x) \in \mathbb{R}^{n}, f: \mathbb{R} \times \mathbb{R}^{n} \rightarrow{\mathbb{R}^{n}}, \end{aligned}$$

where \(f(x,y(x))\) is assumed to be Lipschitz continuous and satisfies conditions for the existence and uniqueness of solutions [8] to the problem (1). Milne [9] is thought to be the inventor of the block methods who employed them to get starting values for the predictor–corrector-type numerical methods. One of the most important research works on the block methods can be seen in [10], proposed way back in 1969. Later, several classical works were reported in [1115] whereas some recently developed implicit block methods include [1625] and most of the references cited therein. Higher-order block methods including one-, two-, and three-step approaches are proposed in the literature with the \(\mathcal{A}\)-stability feature. Still, they are, to some extent, computationally expensive, while some lower-order block methods, although time-efficient, are not accurate enough for stiff systems. In the recent past, Ramos et al. [26] presented an idea for optimization of the intermediate collocation points for k-step hybrid block methods from leading terms of the local truncation errors. Motivated by this research study, several block methods were proposed, including [2729]. For example, Singh et al. [30] proposed a fifth-order block method with three off-step points in a one-step block while optimizing two off-step points and fixing the third, wherein the resulting block method was observed to have \(\mathcal{A}\)-stability. Singh and Ramos [31] designed an optimized seventh-order two-step hybrid block method with two intrastep points for numerical integration of second-order differential systems considering both fixed and variable step-size approaches. Similarly, Ramos [32] presented, while using interpolation and collocation techniques, a fifth-order two-step hybrid block method for solving first-order initial value problems, wherein the method considers minimization of the local truncation errors to optimize the intermediate two off-step points. Several other researches on the block methods can be found in the available scientific literature. Other works are recently reported in [3335] wherein the authors have proposed variable-order techniques, including Haar wavelet collocations, to solve differential equations.

Our new one-step hybrid block technique takes into account three intra-step points and is inspired by recent advancements in the block method field. Minimizing the local truncation error leads to an optimal solution. It is important to note that there is currently no recorded one-step block approach with three intra-step points, of which only one is optimal, in the scientific literature. In this study, we seek to create a new one-step method that, unlike numerous existing block methods, just requires one step to minimize the major term of the local truncation error (see the section of numerical results). Using interpolation and collocation, the suggested method derives a continuous approximation of a polynomial.

The present paper is structured as follows: The proposed implicit hybrid block method is derived in Sect. 2 whereas its theoretical analysis, containing error constant from the local truncation error, consistency, zero-stability, linear stability, and the theory of order stars, is described in Sect. 3. Several physical models are considered in Sect. 4 to test the performance of the proposed method, and the final concluding remarks are presented in Sect. 5.

2 Derivation of the proposed block method

This section presents details about derivation of the proposed one-step hybrid block method with three intrastep points \(x_{n+u}, x_{n+v}\), \(x_{n+w}\) (\(0< u< v< w<1\)) where one point is optimized from the principal term of the local truncation error. The method will later be used to numerically solve several initial value problems as in (1), in both scalar and vector forms. Consider a polynomial approximation of the actual solution \(y(x)\) of (1) at the grid points \(x_{0}< x_{1}< \cdots < x_{N}=T\) over the integration interval \([x_{0}, T]\), with fixed step-size \(\Delta x = x_{k+1}-x_{k}, k = 0,1, \ldots, N-1\). Suppose the following approximate solution:

$$\begin{aligned} y(x) \approx L(x)=\sum_{j=0}^{5} \zeta _{j} x^{j}. \end{aligned}$$

Differentiation of Eq. (2) produces

$$\begin{aligned} y'(x) \approx L'(x)=\sum _{j=1}^{5} j \zeta _{j} x^{j-1}, \end{aligned}$$

where \(\zeta _{j} \in \mathbb{R}\) denote real undetermined coefficients. Consider three intrastep points \(x_{n+u}=x_{n}+u \Delta x, x_{n+v}=x_{n}+v \Delta x,x_{n+w}=x_{n}+w \Delta x\) with \(0< u< v< w<1\) for finding approximate solution of (1) on \([x_{n}, x_{n+1}]\) at the points \(x_{n}\) and \(x_{n+1}\). To carry out the process, consider the approximation in (2) computed at the point \(x_{n}\), and its first-order derivative computed at the points \(x_{n}, x_{n+u}, x_{n+v}, x_{n+w}, x_{n+1}\). This setup brings the following matrix form of a linear system containing six equations in six real unknown coefficients \(\zeta _{j}, j = 0, 1, \ldots, 5\):

$$\begin{aligned} \begin{pmatrix} 1 & x_{n} & x_{n}^{2}&x_{n}^{3}&x_{n}^{4}&x_{n}^{5} \\ 0 & 1 & 2x_{n}&3x_{n}^{2}&4x_{n}^{3}&5x_{n}^{4} \\ 0 & 1 & 2x_{n+u}&3x_{n+u}^{2}&4x_{n+u}^{3}&5x_{n+u}^{4} \\ 0 & 1 & 2x_{n+v}&3x_{n+v}^{2}&4x_{n+v}^{3}&5x_{n+v}^{4} \\ 0 & 1 & 2x_{n+w}&3x_{n+w}^{2}&4x_{n+w}^{3}&5x_{n+w}^{4} \\ 0 & 1 & 2x_{n+1}&3x_{n+1}^{2}&4x_{n+1}^{3}&5x_{n+1}^{4} \end{pmatrix} \begin{pmatrix} \zeta _{0} \\ \zeta _{1} \\ \zeta _{2} \\ \zeta _{3} \\ \zeta _{4} \\ \zeta _{5} \end{pmatrix} = \begin{pmatrix} y_{n} \\ f_{n} \\ f_{n+u} \\ f_{n+v} \\ f_{n+w} \\ f_{n+1} \end{pmatrix} . \end{aligned}$$

Solving the above linear system gives values of the six unknown coefficients \(\zeta _{j}, j = 0, 1, \ldots, 5\) which, for the sake of brevity, are not shown here. Putting these values in (2), while using the change of variable \(x=x_{n} + t \Delta x\), we obtain

$$\begin{aligned} L(x_{n} + t \Delta x) = \zeta _{0} y_{n}+\Delta x (\eta _{0} f_{n}+ \eta _{u} f_{n+u}+\eta _{v} f_{n+v}+\eta _{w} f_{n+w}+\eta _{1} f_{n+1} ), \end{aligned}$$


ζ 0 = 1 , η 0 = t ( 20 u v t 2 30 u v t w 15 u t 3 + 20 u t 2 w 15 v t 3 + 20 v t 2 w + 12 t 4 15 t 3 w 30 u v t + 60 u v w + 20 u t 2 30 u t w + 20 v t 2 ) 60 u v w , η u = t 2 ( 15 v t 2 20 v t w 12 t 3 + 15 t 2 w 20 v t + 30 v w + 15 t 2 20 t w ) 60 ( u 1 ) ( u w ) ( u v ) u , η v = t 2 ( 15 u t 2 20 u t w 12 t 3 + 15 t 2 w 20 u t + 30 u w + 15 t 2 20 t w ) 60 ( v 1 ) ( v w ) ( u v ) v , η w = t 2 ( 20 u v t 15 u t 2 15 v t 2 + 12 t 3 30 u v + 20 u t + 20 v t 15 t 2 ) 60 ( w 1 ) ( v w ) ( u w ) w , η 1 = t 2 ( 20 u v t 30 u v w 15 u t 2 + 20 u t w 15 v t 2 + 20 v t w + 12 t 3 15 t 2 w ) 60 ( w 1 ) ( v 1 ) ( u 1 ) .

Now to get the required one-step hybrid block method, we evaluate \(L(x_{n}+t \Delta x)\) at the collocation points \(x_{n+u}, x_{n+v}\), \(x_{n+w}\), and \(x_{n+1}\) where \(t = u, v, w, 1\). By so doing, the following system of four equations with four unknowns is obtained:

$$\begin{aligned} \begin{aligned} y_{n+u}={}&y_{n} \\ &{} -\frac {\Delta x}{60} \biggl( \frac {-3u^{4}+5u^{3}v+5u^{3}w-10u^{2}vw+5u^{3}-10u^{2}v-10u^{2}w+30uvw}{vw}f_{n} \\ &{} + \frac {u(12u^{3}-15u^{2}v-15u^{2}w+20uvw-15u^{2}+20uv+20uw-30vw)}{(u-1)(u-w)(u-v)}f_{n+u} \\ &{} -\frac {u^{2}(-3u^{3}+5u^{2}w+5u^{2}-10uw)}{(v-1)(v-w)(u-v)v}f_{n+v}- \frac {u^{2}(-3u^{3}+5u^{2}v+5u^{2}-10uv)}{(w-1)(v-w)(u-w)w}f_{n+w} \\ &{} -\frac {u^{2}(-3u^{3}+5u^{2}v+5u^{2}w-10uvw)}{(w-1)(v-1)(u-1)}f_{n+1} \biggr), \end{aligned} \end{aligned}$$
$$\begin{aligned} \begin{aligned} y_{n+v}={}&y_{n} \\ &{} -\frac{\Delta x}{60} \biggl( \frac{5uv^{3}-10uv^{2}w-3v^{4}+5v^{3}w-10uv^{2}+30uvw+5v^{3}-10v^{2}w}{uw}f_{n} \\ &{} +\frac{v^{2}(-3v^{3}+5v^{2}w+5v^{2}-10vw)}{(u-1)(u-w)(u-v)u}f_{n+u} \\ &{} - \frac{v(-15uv^{2}+20uvw+12v^{3}-15v^{2}w+20uv-30uw-15v^{2}+20vw)}{(v-1)(v-w)(u-v)}f_{n+v} \\ &{} +\frac{v^{2}(5uv^{2}-3v^{3}-10uv+5v^{2})}{(w-1)(v-w)(u-w)w}f_{n+w}- \frac{v^{2}(5uv^{2}-10uvw-3v^{3}+5v^{2}w)}{(w-1)(v-1)(u-1)}f_{n+1} \biggr), \end{aligned} \end{aligned}$$
$$\begin{aligned} \begin{aligned} y_{n+w}={}&y_{n} \\ &{} -\frac{\Delta x}{60} \biggl( \frac{-10uvw^{2}+5uw^{3}+5vw^{3}-3w^{4}+30uvw-10uw^{2}-10vw^{2}+5w^{3}}{uv}f_{n} \\ &{} +\frac{w^{2}(5vw^{2}-3w^{3}-10vw+5w^{2})}{(u-1)(u-w)(u-v)u}f_{n+u}- \frac{w^{2}(5uw^{2}-3w^{3}-10uw+5w^{2})}{(v-1)(v-w)(u-v)v}f_{n+v} \\ &{} + \frac{w(20uvw-15uw^{2}-15vw^{2}+12w^{3}-30uv+20uw+20vw-15w^{2})}{(w-1)(v-w)(u-w)}f_{n+w} \\ &{} -\frac{w^{2}(-10uvw+5uw^{2}+5vw^{2}-3w^{3})}{(w-1)(v-1)(u-1)}f_{n+1} \biggr), \end{aligned} \end{aligned}$$
$$\begin{aligned} y_{n+1}={}&y_{n} \\ &{} -\frac{\Delta x}{60} \biggl( \frac{30uvw-10uv-10uw-10vw+5u+5v+5w-3}{uvw}f_{n} \\ &{} +\frac{-10vw+5v+5w-3}{(u-1)(u-w)(u-v)u}f_{n+u}- \frac{-10uw+5u+5w-3}{(v-1)(v-w)(u-v)v}f_{n+v} \\ &{} +\frac{-10uv+5u+5v-3}{(w-1)(v-w)(u-w)w}f_{n+w} \\ &{} - \frac{(-30uvw+20uv+20uw+20vw-15u-15v-15w+12)}{(w-1)(v-1)(u-1)}f_{n+1} \biggr), \end{aligned}$$

where \(y_{n+l} \simeq y(x_{n}+l \Delta x)\) are approximations of the exact solution, and \(f_{n+l}=f(x_{n+l},y_{n+l})\), for \(l=u,v,w,1\). The above approximations consist of the parameters \(u, v, w\) that are related to the intrastep points \(x_{u}, x_{v}, x_{w}\). The suitable values of these parameters can be computed with the help of the principal term of the local truncation error in the main formula, that is, \(y_{n+1}\). One of the parameters will be optimized by imposing a condition on the principal term to vanish. The reason to consider the principal term of the local truncation error in \(y_{n+1}\) to compute one of the optimized parameters is that at the end of the subinterval \([x_{n}, x_{n+1}]\), the value \(y_{n+1}\) is the only value required for advancing the integration on the next block. Thus, to achieve this, we have considered the following local truncation error obtained by expanding \(y(x)\) in the Taylor series about \(x_{n}\). So, the local truncation error in the formula given in (10) is given as follows:

$$\begin{aligned} \begin{aligned} &\mathcal{L}\bigl(y(x_{n+1}); \Delta x\bigr) \\ &\quad= \biggl(\frac {((10w-5)v-5w+3)u}{7200}+\frac {(-5w+3)v}{7200}+ \frac {w}{2400}- \frac {1}{3600} \biggr)\Delta x^{6}y^{(6)}(x_{n}) \\ &\qquad{} + \biggl(\frac {((70w-35)v-35w+21)u^{2}}{302{,}400} \\ &\qquad{} +\frac {(v+w+1)((w-1/2)v-(1/2)w+3/10)u}{4320}+ \frac {(-35w+21)v^{2}}{302{,}400} \\ &\qquad{} +\frac {-35w^{2}-14w+21}{302{,}400}+\frac {w^{2}}{14{,}400}+ \frac {w}{14{,}400}-\frac {1}{12{,}600} \biggr)\Delta x^{7}y^{(7)}(x_{n}) \\ & \qquad{}+\mathcal{O}\bigl(\Delta x^{8}\bigr). \end{aligned} \end{aligned}$$

Equating the principal term (the coefficient of \(\Delta x^{6}\)) of the local truncation error in (11) to zero, we get the following single equation in three unknown parameters:

$$\begin{aligned} \frac {((10w-5)v-5w+3)u}{7200}+\frac {(-5w+3)v}{7200}+\frac {w}{2400}- \frac {1}{3600}=0. \end{aligned}$$

The above equation (12) is expressed in terms of u and v as follows:

$$\begin{aligned} w=\frac {5u v-3v-3u+2}{10uv-5v-5u+3}. \end{aligned}$$

Since there are more unknowns than equations, there exist infinitely many solutions for the parameters under consideration. Considering two of them (say, \(u,v\)) as free parameters, w is optimized. By so doing, the following set of parameters is determined, with w being the optimal one:

$$\begin{aligned} \biggl\{ u = \frac {1}{4}, v = \frac {1}{2}, w = \frac {3}{4} \biggr\} . \end{aligned}$$

Substituting these values into (11), the local truncation error of the main formula in (10) is computed as follows:

$$\begin{aligned} \mathcal{L}\bigl(y(x_{n+1});\Delta x\bigr)=- \frac {(\Delta x)^{7} y^{(7)}(x_{n})}{1{,}935{,}360}+ \mathcal{O}\bigl(\Delta x^{8}\bigr). \end{aligned}$$

Finally, substituting the values of the obtained parameters \(u,v,w\) into (7)–(10) yields the following required one-step hybrid block method with three intrastep points having at least fifth-order convergence while one of the intrastep points is optimized:

$$\begin{aligned} \begin{aligned} &y_{n+\frac{1}{4}}=y_{n}+ \frac {\Delta x}{2880} [251f_{n}+646f_{n+ \frac{1}{4}}-264f_{n+\frac{1}{2}}+106f_{n+\frac{3}{4}}-19f_{n+1} ], \\ &y_{n+\frac{1}{2}}=y_{n}+\frac {\Delta x}{360} [29f_{n}+124f_{n+ \frac{1}{4}}+24f_{n+\frac{1}{2}}+4f_{n+\frac{3}{4}}-f_{n+1} ], \\ &y_{n+\frac{3}{4}}=y_{n}+\frac {\Delta x}{320} [27f_{n}+102f_{n+ \frac{1}{4}}+72f_{n+\frac{1}{2}}+42f_{n+\frac{3}{4}}-3f_{n+1} ], \\ &y_{n+1}=y_{n}+\frac {\Delta x}{90} [7f_{n}+32f_{n+\frac{1}{4}}+12f_{n+ \frac{1}{2}}+32f_{n+\frac{3}{4}}+7f_{n+1} ]. \end{aligned} \end{aligned}$$

The pseudocode for the above proposed method is presented in the algorithm of the Appendix. Moreover, it is easy to notice from the structure of the proposed block method (16) that the slope f occurs four times in each equation without considering \(f_{n}\) as it is just a numerical value. It becomes computationally expensive when f is complicated. This can, however, be overcome by minimizing the number of times f appears in each equation of the above formulae. To accomplish this, we simultaneously solve the set of equations (16) for \(f_{n+u}\), \(f_{n+v}\),, \(f_{n+w}\), and \(f_{n+1}\). It results in an equivalent formulation of (16) reducing the occurrences of f to just one. This proves to be computationally cost-effective, particularly when f is not a simple differentiable function. The obtained reformulation is structured as follows:

$$\begin{aligned} \begin{aligned} &{\Delta x}f_{n+\frac{1}{4}}=- \frac {37}{12}y_{n}+ \frac {2}{3}y_{n+\frac{1}{4}}+3y_{n+\frac{1}{2}}- \frac {2}{3}y_{n+ \frac{3}{4}}+\frac {1}{12}y_{n+1}- \frac {\Delta{x}}{4}f_{n}, \\ &{\Delta x}f_{n+\frac{1}{2}}=\frac {31}{18}y_{n}- \frac {16}{3}y_{n+ \frac{1}{4}}+2y_{n+\frac{1}{2}}+\frac {16}{9}y_{n+\frac{3}{4}}- \frac {1}{6}y_{n+1}+\frac {\Delta{x}}{6}f_{n}, \\ &{\Delta x}f_{n+\frac{3}{4}}=-\frac {29}{12}y_{n}+6y_{n+\frac{1}{4}}-9y_{n+ \frac{1}{2}}+ \frac {14}{3}y_{n+\frac{3}{4}}+\frac {3}{4}y_{n+1}- \frac {\Delta{x}}{4}f_{n}, \\ &{\Delta x}f_{n+1}=\frac {28}{3}y_{n}- \frac {64}{3}y_{n+\frac{1}{4}}+24y_{n+ \frac{1}{2}}-\frac {64}{3}y_{n+\frac{3}{4}}+ \frac {28}{3}y_{n+1}+{ \Delta{x}}f_{n}, \end{aligned} \end{aligned}$$

where the above reformulation of (16) is now abbreviated as \(\mathrm{RPOBM} _{5}\). The function f is required to be evaluated at four points in both formulations; the computational cost, however, is reduced (as evident in the numerical experiments) while using (17) for solving problems as in (1).

3 Theoretical analysis

In this section, we investigate theoretical properties for the proposed implicit hybrid block method given in (16), or equivalently (17), including accuracy, consistency, zero-stability, convergence, linear stability, and \(\mathcal{A}\)-acceptability.

3.1 Order of accuracy and consistency

The hybrid block method in (16) can be rewritten in the following convenient form:

$$\begin{aligned} A_{1} Y_{n+1}=A_{0} Y_{n}+\Delta x(B_{0} F_{n}+B_{1} F_{n+1}), \end{aligned}$$

where \(A_{0},A_{1},B_{0}\), and \(B_{1}\) are \(4 \times 4\) matrices given by

$$\begin{aligned} & A_{0}= \begin{bmatrix} 0&0&0&1 \\ 0&0&0&1 \\ 0&0&0&1 \\ 0&0&0&1 \end{bmatrix},\qquad A_{1} = \begin{bmatrix} 1&0&0&0 \\ 0&1&0&0 \\ 0&0&1&0 \\ 0&0&0&1 \end{bmatrix}, \end{aligned}$$
$$\begin{aligned} &B_{0}= \begin{bmatrix} 0&0&0&\frac{251}{2880} \\ 0&0&0&\frac{29}{360} \\ 0&0&0&\frac{27}{320} \\ 0&0&0&\frac{7}{90} \end{bmatrix},\qquad B_{1}= \begin{bmatrix} \frac{646}{2880}&-\frac{264}{2880}&\frac{106}{2880}&-\frac{19}{2880} \\ \frac{124}{360}&\frac{24}{360}&\frac{4}{360}&-\frac{1}{360} \\ \frac{102}{320}&\frac{72}{320}&\frac{42}{320}&-\frac{3}{320} \\ \frac{32}{90}&\frac{12}{90}&\frac{32}{90}&\frac{7}{90} \end{bmatrix}, \end{aligned}$$


$$\begin{aligned} \begin{aligned} &Y_{n}=(y_{n-1+u},y_{n-1+v},y_{n-1+w},y_{n})^{T}, \\ &Y_{n+1}=(y_{n+u},y_{n+v},y_{n+w},y_{n+1})^{T}, \\ &F_{n}=(f_{n-1+u},f_{n-1+v},f_{n-1+w},f_{n})^{T}, \\ &F_{n+1}=(f_{n+u},f_{n+v},f_{n+w},f_{n+1})^{T}. \end{aligned} \end{aligned}$$

The linear functional operator \(\bar{\mathcal{L}}\) associated with (16) can be defined as

$$\begin{aligned} \bar{\mathcal{L}}\bigl[z(x_{n});\Delta x\bigr] = \sum _{k=0,u,v,w,1} \bigl[ \bar{\zeta}_{k} z(x_{n}+k \Delta x)- \Delta x \bar{\eta}_{k} z'(x_{n}+k \Delta x) \bigr], \end{aligned}$$

where \(\bar{\zeta}_{k}\) and \(\bar{\eta}_{k}\) are the column vectors of the matrices \(A_{1}\) and \(A_{0}\), respectively. The term \(z(x)\) is an arbitrary test function that is considered sufficiently differentiable in the interval \([0, T]\). The block method (16) and the corresponding linear difference operator are said to be at least of order r if after expanding the functions \(z(x_{n}+k \Delta x)\) and \(z'(x_{n}+k\Delta x)\) in Taylor series about \(x_{n}\), and collecting the coefficients of Δx, we obtain

$$\begin{aligned} \bar{\mathcal{L}}\bigl[z(x_{n});\Delta x\bigr]=\bar{C}_{0}z(x_{n})+ \bar{C}_{1} \Delta x z'(x_{n})+ \bar{C}_{2} \Delta x ^{2}z''(x_{n})+ \cdots +\bar{C}_{r} \Delta x^{r}z^{(r)}(x_{n})+ \cdots, \end{aligned}$$

with \(\bar{C}_{0}=\bar{C}_{1} = \cdots =\bar{C}_{r} = 0\) and \(\bar{C}_{r+1}\neq 0\). The coefficients \(\bar{C}_{r}\) are vectors and \(\bar{C}_{r+1}\) is known as the vector of error constants. For the proposed hybrid block method (16), we obtain \(\bar{C}_{0}=\bar{C}_{1} = \cdots =\bar{C}_{5} =0\) with the error constant given by

$$\begin{aligned} \bar{C}_{6}= \biggl(\frac {3}{655{,}360},\frac {1}{368{,}640}, \frac {3}{655{,}360},0 \biggr)^{T}. \end{aligned}$$

Thus, it proves that the one-step implicit hybrid block method with three intrastep points, while one of them is optimally obtained, has at least fifth algebraic order of accuracy. In addition, the method (16) does not have an order less than 1; therefore, it is also consistent with the IVP (1) (see the work of Jator in [36]).

3.2 Zero-stability and convergence

An essential crude requirement for a numerical method to be of some use is the requirement called zero-stability. For example, let the IVP given in (1) be asymptotically stable, whereas the need is to prove the stability of the proposed numerical method (16). The notion of zero-stability relates to considering a homogeneous equation \(y'=0\) and its discretized counterpart as given by

$$\begin{aligned} A_{0} Y_{n+1}-A_{1} Y_{n}=0, \end{aligned}$$

where \(A_{0}\) and \(A_{1}\) are given in (19). Now, if the discrete algebraic equation (25) admits solutions that grow in time, then the proposed block method will not be zero-stable and cannot be used in practice. On the other hand, the proposed block method is said to be zero-stable if zeros \(R_{i}\) of the first characteristic polynomial \(\kappa (R) = |z A_{1}-A_{0}|\) fulfill \(|R_{i}| \leq 1\) and for those zeros with \(|R_{i}|=1\) the multiplicity does not exceed 1 [5]. The first characteristic polynomial of the proposed block method (16) is given by

$$\begin{aligned} \kappa (R) = R^{3} (R-1). \end{aligned}$$

Thus, the proposed block method (16) can be considered a zero-stable method. Being both zero-stable and consistent, it deserves to be called a convergent method (check the work of Henrici in [37]).

3.3 Linear stability analysis and order stars

Theorem 3.1

The proposed one-step hybrid block method with one optimal intrastep point given in (16) satisfies every criterion to be \(\mathcal{A}\)-stable.


As far as the concept of zero stability is concerned, it is related to the behavior of the underlying numerical method as the step-size \(\Delta x \rightarrow{0} \). In other situations, however, a different concept of stability is needed from a practical point of view. It is concerned with a numerical method that produces good results for a particular value of \(\Delta x > 0\). Such behavior is known as the linear stability behavior for the numerical method, and it requires applying the method on a linear test problem proposed by Dahlquist [38], namely

$$\begin{aligned} y'(x)=\sigma y(x), \quad\text{with } \operatorname{Re}( \sigma )\textless 0. \end{aligned}$$

It is required to determine the region wherein the approximations obtained under the numerical method reproduce the behavior of the true solution of the test problem (27). After using the proposed block method (16) on the test problem in (27), the recurrence equation given below is obtained:

$$\begin{aligned} Y_{n}=H(z)Y_{n-1}, \end{aligned}$$

where \(H(z)\) denotes the stability matrix given by

$$\begin{aligned} H(z)=(A_{1}-z B_{1})^{-1}(A_{0}+z B_{0}),\quad z = \sigma \Delta x. \end{aligned}$$

The eigenvalues of the stability matrix (29) determine the behavior of the numerical solution. This is the commonly known stability property of the numerical method that uses the spectral radius (supremum among the absolute values of the elements in the spectrum) of \(H(z)\). The region of absolute linear stability \(\mathbb{A}\) is defined by the set [39]

$$\begin{aligned} \mathbb{A}=\bigl\{ z\in \mathbb{C}: \bigl\vert \rho \bigl[H(z)\bigr] \bigr\vert < 1\bigr\} , \end{aligned}$$

and if \(\mathbb{C}^{-} \subseteq \mathbb{A}\), the underlying numerical method is said to be \(\mathcal{A}\)-stable. The spectral radius is easily computed as the following rational function:

$$\begin{aligned} \rho \bigl[H(z)\bigr]= \frac {3z^{4}+50z^{3}+420z^{2}+1920z+3840}{3z^{4}-50z^{3}+420z^{2}-1920z+3840}, \end{aligned}$$

which has modulus less than one in \(\mathbb{C}^{-}\). This completes the proof for the \(\mathcal{A}\)-stability of the proposed block method given in (16). □

Remark 3.2

The graphical illustrations given by Fig. 1 show that the entire left-half complex plane \(\mathbb{C}^{-}\) is included in the stability region of (16). This kind of \(\mathcal{A}\)- stability is further confirmed with the plot of order stars wherein the rational stability function (31) does not have any pole in \(\mathbb{C}^{-}\) as can be seen in Fig. 2.

Figure 1
figure 1

The stability region in 2D and 3D of the proposed one-step block method (16) with three intrastep points, including one optimal

Figure 2
figure 2

Order stars confirming \(\mathcal{A}\)-stability of the proposed one-step block method (16) with three intrastep points, including one optimal point

4 Numerical dynamics with results and discussion

The proposed one-step hybrid block method with three intrasteps, including one optimal point given in (16), is now employed, neither using any starting values nor any predictors, to determine the approximate solution of various single and systems of stiff ordinary differential equations. Keeping in mind that the initial value \(y(x_{0}) = y_{0}\) is known, the system of equations in (16) is simultaneously solved for \(n=0\). The well-known quadratically convergent Newton’s method is sufficient for solving the system to get the value \(y(x_{1})\cong y_{1}\). Later, taking \(y_{1}\) from the previous block as the initial value, the value \(y(x_{2}) \cong y_{2}\) is computed. This procedure continues until the computation of \(x_{N}\) (the last point). The entire length of the integration interval is chosen to be a multiple of \(2 \Delta x\) (\(x_{N}-x_{0} = \varepsilon (2 \Delta x), \varepsilon \in \mathbb{N} \)), as we have selected, for comparison, some two-step methods. Newton’s method under the Find Root command in Mathematica 12.1 with 64-digit working precision arithmetic is used during the implementation of numerical simulations. All numerical simulations are carried out in Mathematica 12.1 on a personal computer running Windows OS with Intel(R) Core(TM) i7-1065G7 CPU @ 1.30 GHz and 1.50 GHz processor having 24.0 GB installed RAM.

Considering the potential of block methods to deal with stiff differential models, we have taken some well-known stiff applied problems that have appeared several times in the recent literature. These problems are solved with methods as given below:

  • Proposed (\(\mathrm{POBM} _{5}\)) one-step fifth-order hybrid block method with three intrastep grid points shown in (16).

  • Reformulated proposed (\(\mathrm{RPOBM} _{5}\)) one-step fifth-order hybrid block method with three intrastep points shown in (17).

  • Block hybrid Simpson’s method (\(\mathrm{BHSM} _{5}\)) with two intrastep points for the stiff system that appeared in [40].

  • Two-step sixth-order block method (\(\mathrm{Sahi} _{6}\)) that appeared in [41].

  • Fully-implicit RK-type fifth-order method, called Radau method (Radau I5), that appeared in [1]. The Radau family is well-known for solving stiff differential systems.

Moreover, the performance of each method under consideration is measured on different types of error, including maximum global absolute errors \(\mathrm{ME}= (\max_{x \in [x_{0}, T]} |y(x_{N}) - y_{N}| )\), absolute error at final grid point \(\mathrm{LE}=(|y(x_{N}) - y_{N}|)\), average absolute error \(\mathrm{AE}=\frac {1}{N} (\sum_{i=1}^{N} | y(x_{i}) - y_{i} | )\), norm \((\sqrt{\sum_{i=1}^{N} | y(x_{i}) - y_{i} |^{2}} )\), and the CPU time computed in seconds.

The problems under consideration are the most challenging IVPs in ODEs, including burning a ball for a match stick (flame propagation), Prothero–Robinson stiff problem, Kaps stiff system, and the nonlinear Blasius third-order ODE commonly encountered in fields like heat transfer and the computational fluid dynamics. The numerical simulations for each numerical experiment are shown in the Tables 15. Different absolute errors and CPU times are noted to compare the above-discussed methods.

Table 1 Comparison of the methods under consideration on the basis of absolute errors and CPU time for the flame propagation model given in Problem 1 with different number of steps (n)
Table 2 Comparison of the methods under consideration on the basis of absolute errors and CPU time for Problem 2 with different number of steps (n)
Table 3 Comparison of the methods under consideration on the basis of absolute errors and CPU time for the stiff Problem 3 with different number of steps (n)
Table 4 Comparison of the methods under consideration on the basis of absolute errors and CPU time for Kaps Problem 4 with different number of steps (n)
Table 5 Comparison of absolute errors at \(x=10\) for the Blasius equation of boundary layer flow given in Problem 5 with different number of steps (n)

It can be observed from these tables that the proposed block method (16) yields minor errors. In contrast, its reformulation also produces minor errors, reducing CPU times as an additional advantage over the proposed block method. When compared to other well-known methods from available literature, it is noted that the errors are more significant in magnitude; however, the CPU times are as promising as in the reformulated block method in some cases, particularly when the method has a simple form of coefficients such as that in [40]. The efficiency curves shown in Figs. 4 and 3 for the first and second numerical experiments, respectively, are obtained wherein the better performance of the proposed block method (16) and its reformulation given in (17) is seen in terms of absolute maximum global error and CPU times. Thus, the reformulated version of the proposed block method is proved to be the most time-efficient algorithm among considered to obtain an approximate solution of the IVP (1).

Figure 3
figure 3

Efficiency curves of the maximum absolute global errors versus CPU times (s) for highly stiff IVP of the flame propagation given in the numerical experiment 1 with tolerance = \(10^{-i}\), \(i=7,9,11\)

Figure 4
figure 4

Efficiency curves of the maximum absolute global errors versus CPU times (s) for highly stiff IVP of the flame propagation given in the numerical experiment 1 with tolerance = \(10^{-i}\), \(i=7,9,11\)

Problem 1

Consider the following highly stiff IVP for the flame propagation taken from [42]:

$$\begin{aligned} u_{1}'(x)=u_{1}^{2}(x)-u_{1}^{3}(x),\quad u_{1}(0)=\epsilon, x \in \biggl[0,\frac {2}{\epsilon} \biggr], \end{aligned}$$

with the exact solution \(u_{1}(x)=\frac {1}{\mathrm{ProductLog} [9\exp{(-x+9)} ]+1}\), where the function ProductLog is a special function whose details are available in the Wolfram language documentation.

Problem 2

We consider the well-known Prothero–Robinson model taken from [43]:

$$\begin{aligned} u_{1}'(x)=\lambda \bigl(u_{1}(x)-f(x) \bigr)+f'(x), \quad u_{1}(0)=0, x \in [0, 5], \end{aligned}$$

with \(\lambda =10^{-7}, f(x)=\sin(x)\), whose exact solution is \(u_{1}(x)=\sin(x)\).

Problem 3

We consider the following stiff system of first-order ODEs taken from [41]:

$$\begin{aligned} \begin{aligned} u_{1}'(x)&=-u_{1}(x)+95v_{1}(x),\quad u_{1}(0)=1, \\ v_{1}'(x)&=-u_{1}(x)-97v_{1}(x),\quad v_{1}(0)=1, \end{aligned} \end{aligned}$$

where \(x \in [0,2]\). The exact solution of the above system is

$$\begin{aligned} \begin{aligned}& u_{1}(x)=\frac {1}{47} \bigl[95 \exp{(-2x)}-48\exp{(-96x)} \bigr], \\ &v_{1}(x)=\frac {1}{47} \bigl[48\exp{(-96x)}-\exp{(-2x)} \bigr]. \end{aligned} \end{aligned}$$

Problem 4

We consider the stiff system of first-order ODEs known as the Kaps problem taken from [31]:

$$\begin{aligned} \begin{aligned}& u_{1}'(x)=-1002u_{1}(x)+1000v_{1}(x),\qquad u_{1}(0)=1, \\ &v_{1}'(x)=u_{1}(x)-v_{1}(x) \bigl(1+v_{1}(x)\bigr),\qquad v_{1}(0)=1, \end{aligned} \end{aligned}$$

where \(x \in [0,5]\). The exact solution of the above system is

$$\begin{aligned} \begin{aligned}& u_{1}(x)=\exp{(-2x)}, \\ &v_{1}(x)=\exp{(-x)}. \end{aligned} \end{aligned}$$

Problem 5

Consider the following nonlinear third-order ODE known as the Blasius equation of boundary layer flow taken from [44]:

$$\begin{aligned} 2 \frac {d^{3}u_{1}(x)}{dv_{1}^{3}} + u_{1}(x) \frac {d^{2}u_{1}(x)}{dv^{2}}=0,\quad u_{1}(0)=0, u_{1}'(0) = 0, u_{1}''(0)=1. \end{aligned}$$

The above nonlinear equation can be rewritten as a system of three first-order ODEs as shown below:

$$\begin{aligned} &u_{1}'(x)=v_{1}(x),\quad u_{1}(0)=0, \\ &v_{1}'(x)=w_{1}(x),\quad v_{1}(0)=0, \\ &w_{1}'(x)=-\frac {1}{2}u_{1}(x)w_{1}(x),\quad w_{1}(0)=1. \end{aligned}$$

The reference solution of (38) accurate to 30 dp is given at the final mesh point \(x=10\) as follows:

$$\begin{aligned} & \begin{pmatrix} u_{1}(x) \\ v_{1}(x) \\ w_{1}(x) \end{pmatrix} \bigg|_{x=10}\\ &\quad = \begin{pmatrix} 18.369111622637133682327485837045780998471090734582 \\ 2.0854091764379035978623786489081898250823577485323 \\ 1.9055567124740834810772164777626634323571010482536 \times 10^{-18} \end{pmatrix} . \end{aligned}$$

5 Concluding remarks

This study presents a new one-step hybrid block approach with three intra-step points, including one optimally determined from the proposed method’s local truncation error. The order of accuracy, consistency, zero-stability, convergence, \(\mathcal{A}\)-stability, and the theory of order stars are all thoroughly discussed. When the suggested technique was tried on certain tough models from applied sciences, the accuracy results were encouraging. Furthermore, the inclusion of an efficiency plot comparing the errors of the proposed and existing numerical approaches in terms of computing cost for an increasing number of grid points supports reformulation of the suggested method. Future research will focus on the fully implicit block approach with three optimal intra-step points. This will lead to a technique that is truly optimal and is easy to rewrite in a computer. The field of fractional calculus [4550] has additional opportunities to supplement the current study. The proposed hybrid block method can be changed so that it can be used to solve initial value problems of fractional order.

Availability of data and materials

Not applicable.

Code availability

Not applicable.


  1. Butcher, J.C.: Numerical Methods for Ordinary Differential Equations, 3rd edn. Wiley, New York (2016)

    Book  MATH  Google Scholar 

  2. Jay, L.O.: Lobatto methods. Chemistry 29, 298–305 (1996)

    Google Scholar 

  3. Olanegan, O.O., Ogunware, B.G., Omole, E.O., Oyinloye, T.S., Enoch, B.T.: Some variable hybrids linear multistep methods for solving first order ordinary differential equations using Taylor’s series. IOSR J. Math. 11(5), 8–13 (2015)

    Google Scholar 

  4. Fatokun, J.O.: A fifth order collocation method for solving first order initial value problems. J. Inst. Math. Comput. Sci. Math. Ser. 17(1), 73–79 (2006)

    MathSciNet  Google Scholar 

  5. Lambert, J.D.: Computational Methods in Ordinary Differential Equations. Wiley, New York (1973)

    MATH  Google Scholar 

  6. Qureshi, S., Ramos, H.: L-stable explicit nonlinear method with constant and variable step-size formulation for solving initial value problems. Int. J. Nonlinear Sci. Numer. Simul. 19(7–8), 741–751 (2018)

    Article  MathSciNet  MATH  Google Scholar 

  7. Fairuz, A.N., Majid, Z.A.: Rational methods for solving first-order initial value problems. Int. J. Comput. Math. 98(2), 252–270 (2021)

    Article  MathSciNet  MATH  Google Scholar 

  8. Chapra, S.C.: Applied Numerical Methods with MATLAB for Engineers and Scientists, pp. 335–359. McGraw-Hill, New York (2008)

    Google Scholar 

  9. Milne, W.E.: Numerical Solution of Differential Equations. Wiley, New York (1953)

    MATH  Google Scholar 

  10. Shampine, L.F., Watts, H.A.: Block implicit one-step methods. Math. Comput. 23(108), 731–740 (1969)

    Article  MathSciNet  MATH  Google Scholar 

  11. Majid, Z.A., Suleiman, M.B.: Implementation of four-point fully implicit block method for solving ordinary differential equations. Appl. Math. Comput. 184(2), 514–522 (2007)

    MathSciNet  MATH  Google Scholar 

  12. Ngwane, F.F., Jator, S.N.: Block hybrid-second derivative method for stiff systems. Int. J. Pure Appl. Math. 80(4), 543–559 (2012)

    MATH  Google Scholar 

  13. Akinfenwa, O.A., Yao, N.M., Jator, S.N.: A self-starting block Adams method for solving stiff ordinary differential equation. In: 2011 14th IEEE International Conference on Computational Science and Engineering, pp. 127–136. IEEE Press, New York (2011)

    Chapter  Google Scholar 

  14. Anake, T.A.: Continuous implicit hybrid one-step methods for the solution of initial value problems of general second-order ordinary differential equations. PhD dissertation, Covenant University, Ota (2011)

  15. Anuar, K.H.K., Othman, K.I., Ibrahim, Z.B.: Derivation of 3-point block method formula for solving first order stiff ordinary differential equations. In: Proceedings of the 4th International Conference on Applied Mathematics, Simulation, Modelling, pp. 61–65. (2010).

    Google Scholar 

  16. Kashkari, B.S., Syam, M.I.: Optimization of one step block method with three hybrid points for solving first-order ordinary differential equations. Results Phys. 12, 592–596 (2019)

    Article  Google Scholar 

  17. Ibrahim, Z.B., Nasarudin, A.A.: A class of hybrid multistep block methods with \(\mathcal{A}\)-stability for the numerical solution of stiff ordinary differential equations. Mathematics 8(6), 914 (2020)

    Article  Google Scholar 

  18. Akinfenwa, O.A., Abdulganiy, R.I., Akinnukawe, B.I., Okunuga, S.A.: Seventh order hybrid block method for solution of first order stiff systems of initial value problems. J. Egypt. Math. Soc. 28(1), 1–11 (2020)

    Article  MathSciNet  MATH  Google Scholar 

  19. Mehrkanoon, S., Majid, Z.A., Suleiman, M.: A variable step implicit block multistep method for solving first-order ODEs. J. Comput. Appl. Math. 233, 2387–2394 (2010)

    Article  MathSciNet  MATH  Google Scholar 

  20. Osa, A.L., Olaoluwa, O.E.: A fifth-fourth continuous block implicit hybrid method for the solution of third order initial value problems in ordinary differential equations. Appl. Comput. Math. 8(3), 50–57 (2019)

    Article  Google Scholar 

  21. Olusola, K.J.: Block methods for direct solution of higher order ordinary differential equations using interpolation and collocation approach. PhD dissertation, Universiti Utara Malaysia (2015)

  22. Akinfenwa, O.A., Ablganiy, R., Okunuga, S.A., Irechukwu, V.: Simpson’s 3/8-type block method for stiff systems of ordinary differential equations. J. Niger. Math. Soc. 36(3), 503–514 (2017)

    MathSciNet  MATH  Google Scholar 

  23. Omole, E.O., Ogunware, B.G.: 3-Point Single Hybrid Block Method (3PSHBM) for direct solution of general second order initial value problem of ordinary differential equations. J. Sci. Res. Rep. 1–11 (2018)

  24. Ramos, H., Qureshi, S., Soomro, A.: Adaptive step-size approach for Simpson’s-type block methods with time efficiency and order stars. Comput. Appl. Math. 40(6), 1–20 (2021)

    Article  MathSciNet  MATH  Google Scholar 

  25. Qureshi, S., Soomro, A., Hınçal, E.: A New Family of \(\mathcal{A}\)-acceptable Nonlinear Methods with Fixed and Variable Stepsize Approach. Comput. Math. Methods e1213 (2021)

  26. Ramos, H., Kalogiratou, Z., Monovasilis, T., Simos, T.E.: An optimized two-step hybrid block method for solving general second order initial-value problems. Numer. Algorithms 72(4), 1089–1102 (2016)

    Article  MathSciNet  MATH  Google Scholar 

  27. Dong, M., Simos, T.E.: A new high algebraic order efficient finite difference method for the solution of the Schrödinger equation. Filomat 31(15), 4999–5012 (2017)

    Article  MathSciNet  MATH  Google Scholar 

  28. Wang, Z., Simos, T.E.: An economical eighth-order method for the approximation of the solution of the Schrödinger equation. J. Math. Chem. 55(3), 717–733 (2017)

    Article  MathSciNet  MATH  Google Scholar 

  29. Simos, T.E., Tsitouras, C.: A new family of 7 stages, eighth-order explicit Numerov-type methods. Math. Methods Appl. Sci. 40(18), 7867–7878 (2017)

    Article  MathSciNet  MATH  Google Scholar 

  30. Singh, G., Garg, A., Kanwar, V., Ramos, H.: An efficient optimized adaptive step-size hybrid block method for integrating differential systems. Appl. Math. Comput. 362, 124567 (2019)

    MathSciNet  MATH  Google Scholar 

  31. Singh, G., Ramos, H.: An optimized two-step hybrid block method formulated in variable step-size mode for integrating \(y ''= f (x, y, y')\) numerically. Numer. Math., Theory Methods Appl. 12, 640–660 (2019)

    Article  MathSciNet  MATH  Google Scholar 

  32. Ramos, H.: Development of a new Runge–Kutta method and its economical implementation. Comput. Math. Methods 1(2), e1016 (2019)

    Article  MathSciNet  Google Scholar 

  33. Shah, K., Naz, H., Sarwar, M., Abdeljawad, T.: On spectral numerical method for variable-order partial differential equations. AIMS Math. 7(6), 10422–10438 (2022)

    Article  MathSciNet  Google Scholar 

  34. Shah, K., Arfan, M., Ullah, A., Al-Mdallal, Q., Ansari, K.J., Abdeljawad, T.: Computational study on the dynamics of fractional order differential equations with applications. Chaos Solitons Fractals 157, 111955 (2022)

    Article  MathSciNet  Google Scholar 

  35. Shah, K., Jarad, F., Abdeljawad, T.: Stable numerical results to a class of time-space fractional partial differential equations via spectral method. J. Adv. Res. 25, 39–48 (2020)

    Article  Google Scholar 

  36. Jator, S.N.: A sixth order linear multistep method for the direct solution of \(y'' = f (x, y, y^{\prime})\). Int. J. Appl. Math. Comput. Sci. 40(4), 457–472 (2007)

    MathSciNet  MATH  Google Scholar 

  37. Henrici, P.: Discrete Variable Methods in Ordinary Differential Equations. Wiley, New York (1962)

    MATH  Google Scholar 

  38. Dahlquist, G.G.: A special stability problem for linear multistep methods. BIT Numer. Math. 3(1), 27–43 (1963)

    Article  MathSciNet  MATH  Google Scholar 

  39. Wanner, G., Hairer, E.: Solving Ordinary Differential Equations II. Springer, Berlin (1996)

    MATH  Google Scholar 

  40. Biala, T.A., Jator, S.N., Adeniyi, R.B., Ndukum, P.L.: Block hybrid Simpson’s method with two offgrid points for stiff system. Int. J. Nonlinear Sci. 20(1), 3–10 (2015)

    MathSciNet  MATH  Google Scholar 

  41. Sahi, R.K., Jator, S.N., Khan, N.A.: A Simpson’s-type second derivative method for stiff systems. Int. J. Pure Appl. Math. 81(4), 619–633 (2012)

    MATH  Google Scholar 

  42. Shampine, L.F.: Numerical Solution of Ordinary Differential Equations. CRC Press, Boca Raton (1994)

    MATH  Google Scholar 

  43. Prothero, A., Robinson, A.: On the stability and accuracy of one-step methods for solving stiff systems of ordinary differential equations. Math. Comput. 28(125), 145–162 (1974)

    Article  MathSciNet  MATH  Google Scholar 

  44. Adesanya, A.O., Udoh, D.M., Ajileye, A.M.: A new hybrid block method for the solution of general third order initial value problems of ordinary differential equations. Int. J. Pure Appl. Math. 86(2), 365–375 (2013)

    Article  Google Scholar 

  45. Rihan, F.A., Al-Mdallal, Q.M., AlSakaji, H.J., Hashish, A.: A fractional-order epidemic model with time-delay and nonlinear incidence rate. Chaos Solitons Fractals 126, 97–105 (2019)

    Article  MathSciNet  MATH  Google Scholar 

  46. Qureshi, S., Abro, K.A., Gómez-Aguilar, J.F.: On the numerical study of fractional and non-fractional model of nonlinear Duffing oscillator: a comparison of integer and non-integer order approaches. Int. J. Model. Simul. 1–14 (2022)

  47. Al-Mdallal, Q.M., Yusuf, H., Ali, A.: A novel algorithm for time-fractional foam drainage equation. Alex. Eng. J. 59(3), 1607–1612 (2020)

    Article  Google Scholar 

  48. Al-Mdallal, Q.M.: On fractional-Legendre spectral Galerkin method for fractional Sturm–Liouville problems. Chaos Solitons Fractals 116, 261–267 (2018)

    Article  MathSciNet  MATH  Google Scholar 

  49. Yusuf, A., Qureshi, S., Mustapha, U.T., Musa, S.S., Sulaiman, T.A.: Fractional modeling for improving scholastic performance of students with optimal control. Int. J. Appl. Comput. Math. 8(1), 1–20 (2022)

    Article  MathSciNet  Google Scholar 

  50. Al-Mdallal, Q.M., Syam, M.I., Anwar, M.N.: A collocation-shooting method for solving fractional boundary value problems. Commun. Nonlinear Sci. Numer. Simul. 15(12), 3814–3822 (2010)

    Article  MathSciNet  MATH  Google Scholar 

Download references


Asifa Tassaddiq would like to thank the Deanship of Scientific Research at Majmaah University for supporting this work under Project No. R-2022-325. The authors are also thankful to the worthy reviewers and editors for their useful and valuable suggestions for the improvement of this paper which led to a better presentation.


No specific external funding was received for this research work.

Author information

Authors and Affiliations



The first, second and third authors have conceived the idea and derived the proposed optimized block method. The first and second author has supervised the entire work and guided, whereas the fourth author carried out a theoretical analysis of the proposed method. Finally, the fifth author performed numerical simulations. Each author equally contributed towards writing and finalizing the article.

Corresponding author

Correspondence to Asifa Tassaddiq.

Ethics declarations

Ethics approval and consent to participate

We with this affirm that the contents of this article are original. Furthermore, it has been neither published elsewhere fully or partially in any language nor submitted for publication (fully or partially) elsewhere simultaneously. It contains no matter that is scandalous, obscene, fraud, plagiarism, libelous, or otherwise contrary to law.

Consent to participate

Each author has approved of and agreed to submit the article.

Consent for publication

Each author agreed to publish the article.

Competing interests

The authors declare no competing interests.



Algorithm 1
figure a

Pseudocode for the optimal one-step \(\mathcal{A}\)-stable block method

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Tassaddiq, A., Qureshi, S., Soomro, A. et al. A new continuous hybrid block method with one optimal intrastep point through interpolation and collocation. Fixed Point Theory Algorithms Sci Eng 2022, 22 (2022).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: